Appreciation post: I chose a local model over GPT-4 for my project's final layer

Everyone on my team said it was a dumb move, but running Llama 3.1 locally gave me the control to fine-tune it exactly for our niche medical data. Anyone else get pushback for picking a less popular tool that just fit better?

2 comments

2 Comments

lunaw7226d agoMost Upvoted

Oh man, this reminds me of my friend who was building this little farm supply tracker. He went with some basic open source thing instead of a big cloud service everyone told him to use. His boss was mad at first, but then their internet went out for a week after a storm and his whole team could still work offline. The fancy option would have been totally dead.

kai46326d ago

My buddy's startup insisted on using this obscure database for their mapping app. The big names said it couldn't handle the load, but it processed their custom geo-data twice as fast. Sometimes the right tool is just the one that fits the problem, not the one with the biggest logo.